> ## Documentation Index
> Fetch the complete documentation index at: https://docs.avigrah.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Integrations

Integrations in Avigrah are **system connectors** that enable AI agents and pipelines to interact with external platforms.

They are used for:

* Ingesting data into pipelines
* Executing actions through AI agents

***

## Role in the System

Integrations operate in two primary ways:

### 1. Data Source Layer (Pipelines)

Integrations act as **input sources** for data pipelines.

* Fetch structured data from external systems
* Support batch (API) and real-time (webhook) ingestion
* Feed data into lakehouse and vector systems

This enables pipelines to process real business data.

***

### 2. Execution Layer (Agents)

Integrations are directly used by AI agents to **perform actions**.

Agents can:

* Create and manage campaigns
* Add or update users
* Trigger workflows in external systems
* Execute platform-specific operations

This moves AI from analysis to **action execution**.

***

## Integration Model

Each integration represents a **provider-specific service connector**.

Examples include:

* CRM systems
* Marketing platforms
* Advertising platforms
* Commerce systems
* Communication tools

Each connector exposes:

* Structured data access
* Actionable endpoints
* Platform-specific capabilities

***

## System Flow

Integrations participate in both data and execution flows:

**Integration → Pipeline → Storage → AI → Integration Action**

* Data is pulled into pipelines
* Processed and structured
* Used by AI agents for reasoning
* Agents trigger actions back through integrations

***

## Project Scope

* Integrations are configured at the organization level
* Assigned to specific projects
* Accessible only within project scope

This ensures controlled and secure usage.

***

## Execution Behavior

When used by agents:

* Requests are routed through integration services
* Actions are executed on external platforms
* Responses are returned to the system

All interactions are governed by permissions and project scope.

***

## Positioning

Integrations are not passive connectors.

They are **active system components** that enable:

* Data ingestion
* External system interaction
* AI-driven execution

They allow Avigrah to function as an **actionable AI system**, not just an analytical layer.

***

## Supported Integrations

<Card title="Shopify" icon="shopify" horizontal>
  Manage store data, orders, and customers while enabling AI-driven commerce actions.
</Card>

<Card title="WooCommerce" icon="bag-shopping" horizontal>
  Handle products, orders, and customer data for AI-powered commerce operations.
</Card>

<Card title="Salesforce" icon="salesforce" horizontal>
  Work with leads, accounts, and pipelines while enabling AI-driven sales actions.
</Card>

<Card title="Zoho CRM" icon="database" horizontal>
  Access CRM records, manage leads, and trigger sales-related workflows.
</Card>

<Card title="Klaviyo" icon="mailbox-flag-up" horizontal>
  Manage campaigns, segments, and customer engagement through AI workflows.
</Card>

<Card title="Mailchimp" icon="mailchimp" horizontal>
  Create campaigns, manage audiences, and automate email marketing actions.
</Card>

<Card title="Meta Ads" icon="meta" horizontal>
  Analyze and manage ad campaigns, audiences, and performance data
</Card>

<Card title="Google Ads" icon="google" horizontal>
  Monitor campaigns, optimize ads, and work with performance data
</Card>

Some integrations may have limited functionality during testing or may be available only for enterprise clients; please contact support for access and details.